On Wed, Jan 10, 2001 at 09:07:21AM -0500, Donald Becker wrote: > That means that the other end was in forced-full-duplex mode all along. > That's a network configuration problem, not a driver problem. I've been wondering about that. However, the other end shouldn't be in forced-full-duplex mode. The other end is an rtl8139 based card, a DFE-538TX, which is not compiled as a module, it's right in the kernel (2.2.18), and the card came up full duplex (according to rtl8139-diag). Would the kernel have done a forced-full-duplex? Or did it autonegotiate properly? The client and server are running through a D-Link DSS+-8 10/100 switch, which I hope doesn't have broken autonegotiation, especially with other D-Link hardware. Although, I stand to be corrected. :) There's another client on the network that's got a 3c905 adapter and vortex-diag reports that it's running in full-duplex mode as well, as a module, with no options passed. Could it be something else? -- Adrian Chung - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org
